[dsn_klr650] missing oil filter tube - back to the original que

[Dreas Nielsen]

I had both a bicycle computer (SigmaSport BC800) and a GPS (Garmin eMap) on the KLR during my Sunday ride, and found the relative accuracy interesting. The GPS usually indicated a speed that was 1.5 to 2 mph faster than the BC, but the BC only reported speed to the nearest single digit, so I couldn't make a precise and consistent comparison. Over the course of the entire ride, however, the GPS indicated a mileage of 491 and the BC a mileage of 485, for a relative percent difference of 1.2. What is interesting about that is that it is a little too good. I calibrated the BC by taping off the distance between two chalk marks on the concrete, equivalent to two revolutions of the front tire. This total distance was approximately 82.9 inches, and I figure I couldn't measure it accurately to nearer than an eighth of an inch, for a relative error of 1.5 percent. I don't know the relative precision of the GPS's distance measurements (neither the Garmin manual nor Web site provides this information) (and it's probably speed-dependent), but I'd be surprised if it was much better than a couple of percent. So I figure the RMS error of the BC and GPS combined should be about 2.5 percent, or roughly twice what I observed. On the other had, if the GPS has a negligible error, the accuracy of the BC is 20% better than I expected, or pretty good (I think) for a relatively crude calibration procedure. So to really figure out what's up, I need 99 others of you out there to do the same comparison, and then we can see if the relative accuracy I found really was a two-sigma event. Get riding! Dreas

[Steve Anderson]

GPS speed error has nothing to do with your speed over the ground. Remember the GPS is calculating your speed relative to satellites whizzing around at unholy speeds, not relative to the earth, which is what you're talking about. Everything in GPS is calculated in terms of differences in position and velocity, by measuring differences in time, with nothing really being "absolute", except for time (GPS's get synchronized to the atomic clocks in the satellites, and are one of the best and most accurate sources of "absolute time" short of getting yourself an atomic clock!). Garmin specifies the speed of the GPSIII(and +) at 0.1 knot, pretty darn good. The errors in the GPS come from the fact that most of the units we use update once per second, and compute speed in terms of the straight line between the two points, even though in a turn you travel a larger distance. In a straight line, the GPS is pretty much dead on. It will read a bit slow in a turn. In actuality, since GPS works on differences, it is actually very good at distance measurements between two points, with any satellite errors averaging completely out to be no more than the basic error. But then again, there is that turning issue. The more you turn the lower it will read compared to what you did. Bottom line: The only real significant error in your GPS is probably due to what they call "straight line interpolation", i.e. figuring you traveled in a perfectly straight line between samples, so the error will be that it reads low, with the error being in proportion to how "not straight" you went. Steve A.

--- In DSN_klr650@egroups.com, "Dreas Nielsen" wrote: > I had both a bicycle computer (SigmaSport BC800) and a GPS (Garmin eMap) on > the KLR during my Sunday ride, and found the relative accuracy interesting. > The GPS usually indicated a speed that was 1.5 to 2 mph faster than the BC, > but the BC only reported speed to the nearest single digit, so I couldn't > make a precise and consistent comparison. Over the course of the entire > ride, however, the GPS indicated a mileage of 491 and the BC a mileage of > 485, for a relative percent difference of 1.2. > > What is interesting about that is that it is a little too good. I > calibrated the BC by taping off the distance between two chalk marks on the > concrete, equivalent to two revolutions of the front tire. This total > distance was approximately 82.9 inches, and I figure I couldn't measure it > accurately to nearer than an eighth of an inch, for a relative error of 1.5 > percent. I don't know the relative precision of the GPS's distance > measurements (neither the Garmin manual nor Web site provides this > information) (and it's probably speed-dependent), but I'd be surprised if it > was much better than a couple of percent. So I figure the RMS error of the > BC and GPS combined should be about 2.5 percent, or roughly twice what I > observed. > > On the other had, if the GPS has a negligible error, the accuracy of the BC > is 20% better than I expected, or pretty good (I think) for a relatively > crude calibration procedure. > > So to really figure out what's up, I need 99 others of you out there to do > the same comparison, and then we can see if the relative accuracy I found > really was a two-sigma event. Get riding! > > Dreas

[Ted Palmer]

Daniel Copeland wrote: [...]

> > 1) What damage may have happened/should I be on the look out for > > after running the KLR from 10,000km to 15,000km (6,000 - 9,000 > miles) > > without any filtering?

The motor would have been just about run in by 10,000km, so most of the little bits of metal would have been collected already. Hard to say how much damage was done. The engine might lose a bit of useful life eventually. Running low on oil on the highway costs more, I know.

> > 2) What can I/should I expect the dealership to do about this?

There's a big difference between "can" and "should". Most likely the dealer will disavow any knowledge of their actions. Their credibility will self-destruct in 5 seconds. Mister_T